中文摘要：

现有图像降维方法中特征信息被过多压缩,从而影响图像分类效果。提出IC-ACO算法,利用蚁群算法来解决图像分类问题。算法充分提取并保留图像的各种形态特征。利用蚁群优化算法在特征集中自动挖掘有效特征和特征值,构建各类分类规则,从而实现图像的分类识别。在真实的车标图像数据集上的实验结果表明,IC-ACO算法比其他类似算法具有更高的分类识别率。

英文摘要：

Feature information in current image dimension reduction methods has been excessively compressed,which impacts the efficiency of image classification.In this paper we present the IC-ACO algorithm,it employs ant colony optimisation to solve image classification problem.The algorithm fully extracts various morphological features of image and retains them.The ant colony optimisation is used to automatically mine effective features and feature values from feature sets,the algorithm then constructs the classification rules of every type,thus realises image’s classified recognition.Experimental results on actual vehicle-logo image data sets show that the IC-ACO algorithm outperforms other similar algorithms in terms of the classified recognition accuracy.